ABSTRACT Ischemic stroke is one of the leading causes of morbidity and mortality in the U.S. The goal of acute ischemic stroke therapy is to salvage tissue that is at risk of infarction, but still viable, through the use of reperfusion strategies. Current reperfusion therapies are limited by a tight time window for treatment and by the potential risk of brain hemorrhage. Using this time-based approach, only a limited number of stroke patients are eligible for treatment. Patients who present beyond the standard treatment time windows can benefit from therapy when identified based on multimodal MRI; however, precise and accurate identification of the salvageable tissue is essential, as the potential beneficial effect of treatment must be weighed against the risk of hemorrhage. Although a diffusion/perfusion MRI mismatch has been suggested as a guide with which to identify the presence of salvageable tissues and to serve as a selection marker for thrombolysis, the results of clinical trials using this criterion have been inconclusive, in part because of the inclusion of regions of oligemia in the penumbra, which overestimates the size of the tissue at risk. Amide proton transfer (APT) MRI has shown promise in detecting such an acidosis-based ischemic penumbra in animal models and in human stroke patients. However, most currently used APT imaging protocols are not very practical and not optimized with respect to the magnitude of signal changes caused by the pH effect. More quantitative APT-MRI typically would require an even longer scan time due to the use of multiple RF saturation frequencies, multiple acquisitions, and a long RF saturation pulse (or pulse train), all of which hamper clinical translation due to the very small time-window between stroke onset and possible thrombolysis treatment. Our long-term goal is to develop an ultrafast pH imaging technique for routine clinical use to guide reperfusion therapies for hyperacute stroke patients at various therapeutic time windows, as well as predict the risk of hemorrhagic transformation (HT) following acute ischemic stroke. The first clinical hypothesis is that, similar to animal studies, the pH imaging penumbra due to ischemic tissue acidosis predicts the maximum final infarction size if no reperfusion is initiated. Our second clinical hypothesis is that the presence of severe tissue acidosis in the ischemic core is associated with an increased probability of secondary HT. Our hypotheses will be tested through three specific aims: 1) to develop and optimize an ultrafast quantitative pH imaging method; 2) to validate this technique and assess the diagnostic accuracy of the acidosis-based ischemic penumbra in a clinical setting; and 3) to develop a novel deep-learning model with which to predict HT following acute ischemic stroke, and quantify the sensitivity and specificity of pH imaging. This work is expected to accelerate the translation of APT-MRI into a clinically viable and robust method. The addition of pH imaging to the standard MRI protocol is expected to enable better visualization of the true ischemic penumbra, thus improving predictions of clinical outcome and reducing the incidence of HT.